diff options
author | Samuel Just <sam.just@inktank.com> | 2013-04-15 16:33:48 -0700 |
---|---|---|
committer | Samuel Just <sam.just@inktank.com> | 2013-05-31 13:36:07 -0700 |
commit | a5f6d8967fbbe87f34b23cfcceea84317b647f62 (patch) | |
tree | 4719b7d9e7781234987dc5b140197b277b8e4944 | |
parent | 9aaff1492a00e137f84af9b03e3a4608cea4b520 (diff) | |
download | ceph-a5f6d8967fbbe87f34b23cfcceea84317b647f62.tar.gz |
PG: don't write out pg map epoch every handle_activate_map
We don't actually need to write out the pg map epoch on every
activate_map as long as:
a) the osd does not trim past the oldest pg map persisted
b) the pg does update the persisted map epoch from time
to time.
To that end, we now keep a reference to the last map persisted.
The OSD already does not trim past the oldest live OSDMapRef.
Second, handle_activate_map will trim if the difference between
the current map and the last_persisted_map is large enough.
Fixes: #4731
Signed-off-by: Samuel Just <sam.just@inktank.com>
Reviewed-by: Greg Farnum <greg@inktank.com>
(cherry picked from commit 2c5a9f0e178843e7ed514708bab137def840ab89)
Conflicts:
src/common/config_opts.h
src/osd/PG.cc
- last_persisted_osdmap_ref gets set in the non-static
PG::write_info
-rw-r--r-- | src/common/config_opts.h | 4 | ||||
-rw-r--r-- | src/osd/PG.cc | 15 | ||||
-rw-r--r-- | src/osd/PG.h | 1 |
3 files changed, 18 insertions, 2 deletions
diff --git a/src/common/config_opts.h b/src/common/config_opts.h index bc94acc4de8..ef00babe68a 100644 --- a/src/common/config_opts.h +++ b/src/common/config_opts.h @@ -430,6 +430,10 @@ OPTION(osd_use_stale_snap, OPT_BOOL, false) OPTION(osd_rollback_to_cluster_snap, OPT_STR, "") OPTION(osd_default_notify_timeout, OPT_U32, 30) // default notify timeout in seconds OPTION(osd_kill_backfill_at, OPT_INT, 0) + +// Bounds how infrequently a new map epoch will be persisted for a pg +OPTION(osd_pg_epoch_persisted_max_stale, OPT_U32, 200) + OPTION(osd_min_pg_log_entries, OPT_U32, 3000) // number of entries to keep in the pg log when trimming it OPTION(osd_max_pg_log_entries, OPT_U32, 10000) // max entries, say when degraded, before we trim OPTION(osd_op_complaint_time, OPT_FLOAT, 30) // how many seconds old makes an op complaint-worthy diff --git a/src/osd/PG.cc b/src/osd/PG.cc index ecca34d50d6..945ea637c6c 100644 --- a/src/osd/PG.cc +++ b/src/osd/PG.cc @@ -139,7 +139,7 @@ PG::PG(OSDService *o, OSDMapRef curmap, p.m_seed, p.get_split_bits(curmap->get_pg_num(_pool.id)), _pool.id), - osdmap_ref(curmap), pool(_pool), + osdmap_ref(curmap), last_persisted_osdmap_ref(curmap), pool(_pool), _lock("PG::_lock"), ref(0), #ifdef PG_DEBUG_REFS @@ -2629,6 +2629,7 @@ void PG::write_info(ObjectStore::Transaction& t) past_intervals, snap_collections, osd->infos_oid, info_struct_v, dirty_big_info); assert(ret == 0); + last_persisted_osdmap_ref = osdmap_ref; dirty_info = false; dirty_big_info = false; @@ -5536,7 +5537,17 @@ void PG::handle_activate_map(RecoveryCtx *rctx) dout(10) << "handle_activate_map " << dendl; ActMap evt; recovery_state.handle_event(evt, rctx); - dirty_info = true; + if (osdmap_ref->get_epoch() - last_persisted_osdmap_ref->get_epoch() > + g_conf->osd_pg_epoch_persisted_max_stale) { + dout(20) << __func__ << ": Dirtying info: last_persisted is " + << last_persisted_osdmap_ref->get_epoch() + << " while current is " << osdmap_ref->get_epoch() << dendl; + dirty_info = true; + } else { + dout(20) << __func__ << ": Not dirtying info: last_persisted is " + << last_persisted_osdmap_ref->get_epoch() + << " while current is " << osdmap_ref->get_epoch() << dendl; + } } void PG::handle_loaded(RecoveryCtx *rctx) diff --git a/src/osd/PG.h b/src/osd/PG.h index fe898791c0c..676601121b2 100644 --- a/src/osd/PG.h +++ b/src/osd/PG.h @@ -383,6 +383,7 @@ public: } protected: OSDMapRef osdmap_ref; + OSDMapRef last_persisted_osdmap_ref; PGPool pool; OSDMapRef get_osdmap() const { |